Automated MV markerless tumor tracking for VMAT.

Area of Science:

• Medical Physics
• Radiotherapy Technology
• Image-Guided Interventions

Background:

• Accurate tumor tracking in radiotherapy is crucial for dose accuracy, conformity, and sparing healthy tissues.
• Evolving treatment complexities (VMAT, IMRT) and hypofractionation trends necessitate advanced motion management strategies.
• Existing methods, while useful, face challenges with dynamic treatment deliveries.

Purpose of the Study:

• To develop and evaluate a novel markerless lung tumor tracking algorithm for radiotherapy.
• To adapt the algorithm for complex treatment deliveries involving gantry and multi-leaf collimator (MLC) motion.
• To assess the algorithm's performance using phantom and patient data during dynamic treatment arcs.

Main Methods:

• Developed a markerless lung tumor tracking algorithm utilizing the electronic portal imager (EPID).
• Employed an adaptive template matching core with stability metrics and relative orientation for simultaneous multi-feature tracking.
• Algorithm initialization requires a single image, with dynamic feature management based on input.

Main Results:

• The algorithm demonstrated robust performance on a dynamic thorax phantom and patient data during VMAT arcs.
• Reported tracking errors were 1.34 mm for phantom data and 0.68 mm for patient data.
• The multi-region, markerless approach achieved robust target localization without a priori information.

Conclusions:

• A novel multi-region, markerless tracking algorithm has been successfully developed for radiotherapy.
• The algorithm provides robust target localization during complex treatments, including VMAT and IMRT.
• The achieved tracking accuracy is comparable to methods used in simpler 3D conformal treatments.